Merge pull request oobabooga#7366 from oobabooga/dev

Merge dev branch

Author: oobabooga

## Chat Customization Diagnostics.md

@@ -0,0 +1,43 @@
+## Chat Customization Diagnostics
+*WARNING: This file may contain sensitive information.*
+
+**Custom Agents**<br>
+*1 file loaded*
+
+.github/agents<br>
+.claude/agents<br>
+User Data<br>
+Extension: GitHub.copilot-chat<br>
+└─ [`Plan.agent.md`](../../../AppData/Roaming/Code/User/globalStorage/github.copilot-chat/plan-agent/Plan.agent.md)<br>
+
+**Instructions**<br>
+
+.github/instructions<br>
+.claude/rules<br>
+~/.claude/rules<br>
+User Data<br>
+
+
+**Prompt Files**<br>
+*3 files loaded*
+
+.github/prompts<br>
+User Data<br>
+Extension: GitHub.copilot-chat<br>
+├─ [`savePrompt.prompt.md`](../../../.vscode/extensions/github.copilot-chat-0.37.6/assets/prompts/savePrompt.prompt.md)<br>
+├─ [`plan.prompt.md`](../../../.vscode/extensions/github.copilot-chat-0.37.6/assets/prompts/plan.prompt.md)<br>
+└─ [`init.prompt.md`](../../../.vscode/extensions/github.copilot-chat-0.37.6/assets/prompts/init.prompt.md)<br>
+
+**Skills**<br>
+
+.github/skills<br>
+.agents/skills<br>
+.claude/skills<br>
+~/.copilot/skills<br>
+~/.agents/skills<br>
+~/.claude/skills<br>
+
+**Hooks**<br>
+
+*No files loaded*
+

.gitignore

@@ -1,4 +1,4 @@
-/css
+$env:OPENAI_API_KEY='sk-REPLACE_WITH_YOUR_KEY'/css
 /extensions
 /installer_files
 /repositories
@@ -31,3 +31,4 @@ wandb
 /docker-compose.yml
 /Dockerfile
 .env
+installer_files/conda/python.exe

css/html_instruct_style.css

@@ -19,12 +19,14 @@
     color: #d1d5db !important;
 }
 
-.chat .message-body :is(th, td) {
+.chat .message-body :is(th, td),
+.prose hr {
     border-color: #40404096 !important;
 }
 
-.dark .chat .message-body :is(th, td) {
-    border-color: #ffffff75 !important;
+.dark .chat .message-body :is(th, td),
+.dark .prose hr {
+    border-color: rgb(255 255 255 / 30%) !important;
 }
 
 .chat .message-body :is(p, ul, ol) {

css/main.css

@@ -1797,3 +1797,20 @@ button#swap-height-width {
     top: 0;
     left: calc(100% - 174px);
 }
+
+table {
+    border-collapse: collapse;
+}
+
+table, tr, td, th, thead {
+    border: 0;
+}
+
+td + td,
+th + th { border-left: 1px solid; }
+
+tr + tr td,
+tr + tr th { border-top: 1px solid; }
+
+thead + tbody tr:first-child td,
+thead + tbody tr:first-child th { border-top: 1px solid; }

download-model.py

@@ -435,3 +435,6 @@ def check_model_files(self, model, branch, links, sha256, output_folder):
             model, branch, links, sha256, output_folder,
             specific_file=specific_file, threads=args.threads, is_llamacpp=is_llamacpp
         )
+
+
+

extensions/whisper_stt/requirements.txt

@@ -1,3 +1,4 @@
+
 SpeechRecognition==3.10.0
 openai-whisper
 soundfile

extensions/whisper_stt/script.js

@@ -62,7 +62,8 @@ window.startStopRecording = function() {
   }
 };
 
-const recordButton = gradioApp().querySelector("#record-button");
+const 
+recordButton = gradioApp().querySelector("#record-button");
 recordButton.addEventListener("click", window.startStopRecording);
 
 
@@ -84,3 +85,5 @@ recButton.innerHTML = "Rec.";
 recButton.addEventListener("click", function() {
   recordButton.click();
 });
+
+'package.json'

modules/models_settings.py

@@ -83,7 +83,11 @@ def get_model_metadata(model):
 
         if 'tokenizer.chat_template' in metadata:
             template = metadata['tokenizer.chat_template']
-            eos_token = metadata['tokenizer.ggml.tokens'][metadata['tokenizer.ggml.eos_token_id']]
+            if 'tokenizer.ggml.eos_token_id' in metadata:
+                eos_token = metadata['tokenizer.ggml.tokens'][metadata['tokenizer.ggml.eos_token_id']]
+            else:
+                eos_token = ""
+
             if 'tokenizer.ggml.bos_token_id' in metadata:
                 bos_token = metadata['tokenizer.ggml.tokens'][metadata['tokenizer.ggml.bos_token_id']]
             else:

one_click.py

@@ -77,6 +77,7 @@ def cpu_has_avx2():
 def cpu_has_amx():
     try:
         import cpuinfo
+        
         info = cpuinfo.get_cpu_info()
         return 'amx' in info['flags']
     except:

requirements/full/requirements.txt

@@ -1,12 +1,12 @@
 accelerate==1.8.*
 audioop-lts<1.0; python_version >= "3.13"
-bitsandbytes==0.48.*
+bitsandbytes==0.49.*
 colorama
 datasets
 diffusers==0.36.*
 einops
 fastapi==0.112.4
-flash-linear-attention==0.4.0
+flash-linear-attention==0.4.*
 html2text==2025.4.15
 huggingface-hub==0.36.0
 jinja2==3.1.6
@@ -26,7 +26,7 @@ safetensors==0.7.*
 scipy
 sentencepiece
 tensorboard
-torchao==0.14.*
+torchao==0.15.*
 transformers==4.57.*
 triton-windows==3.5.1.post22; platform_system == "Windows"
 tqdm
@@ -42,8 +42,8 @@ sse-starlette==1.6.5
 tiktoken
 
 # CUDA wheels
-https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.71.0/llama_cpp_binaries-0.71.0+cu124-py3-none-win_amd64.whl; platform_system == "Windows"
-https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.71.0/llama_cpp_binaries-0.71.0+cu124-py3-none-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64"
+https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.74.0/llama_cpp_binaries-0.74.0+cu124-py3-none-win_amd64.whl; platform_system == "Windows"
+https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.74.0/llama_cpp_binaries-0.74.0+cu124-py3-none-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64"
 https://github.com/turboderp-org/exllamav3/releases/download/v0.0.18/exllamav3-0.0.18+cu128.torch2.7.0-cp311-cp311-win_amd64.whl; platform_system == "Windows" and python_version == "3.11"
 https://github.com/turboderp-org/exllamav3/releases/download/v0.0.18/exllamav3-0.0.18+cu128.torch2.7.0-cp311-cp311-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64" and python_version == "3.11"
 https://github.com/turboderp-org/exllamav2/releases/download/v0.3.2/exllamav2-0.3.2+cu128.torch2.7.0-cp311-cp311-win_amd64.whl; platform_system == "Windows" and python_version == "3.11"